一种改进的自适应线性网络多用户检测器  

An Improved Multi-User Detector Based on Adaline Neural Network

作　　者：王杰令[1] 刘祖军[1] 易克初[1] 胡小林[1] 

机构地区：[1]西安电子科技大学ISN国家重点实验室,陕西西安710071

出　　处：《信号处理》2008年第4期659-662,共4页Journal of Signal Processing

基　　金：国家自然科学基金(60572148);中国博士后科学基金(20070411119)

摘　　要：采用粒子群优化算法结合LMS算法来改进自适应线性网络的训练过程,改变了传统神经网络交错反复周期性循环训练样本的训练方式,可以逐个样本的进行训练来达到全局最优,从而从根本上克服了神经网络动态适应性能差的缺点。计算机仿真结果表明,改进的自适应线性网络构成的多用户检测器(简称ANN-MUD),动态适应性能明显改善,实用性大大增强。The particle swarm optimization （PSO） algorithm combined with LMS algorithm is applied to the training of the adaline Neural Network ,which can effectively remove the shortcoming of poor dynamic adaptive behavior of conventional Neural Network, i. e. in the procedure of training weights of conventional Neural Network,it usually requires the training samples trained iteratively. But in the improved one, each training sample can be trained for many times. Simulation experimental results show that, in the Adaline Neural Network applied to Multi-User Detection improved by the PSO, the dynamic adaptive behavior has been remarkably improved and its practical value increases significantly.

关 键 词：多址干扰 多用户检测 自适应线性网络 粒子群优化 误码率 

分 类 号：TN929.533[电子电信—通信与信息系统] TP273.2[电子电信—信息与通信工程]